
Notes about Bassett ATCL Elicitation
====================================
Conversation took place on 9 Nov 2015

The questions I posed to Troy Bassett included the following:

For $y \in \{1894, 1891, 1886\} X, Y, Z\}$,

0. In the year $y$ *At the Circulating Library* shows that at least $n_y$
novels were published in that year. Consider for the moment, how many *more*
novels published in this year you think might be found given a great deal of
time and resources. There may be a great deal of uncertainty about this number
-- for example, you might feel that it's possible that the number of novels
which would eventually be added likely lies between 0 and 500.

1. Could you settle on a value such that you feel that it is equally likely
that the number of additional novels found in $y$ is equally likely to be
greater as to be less than this value. [Call the answer, the median estimate, $M$].

2. Suppose you were told that the number of additional novels ($n_y$) found in
$y$ is greater than $M$; can you now determine a new value such that you feel
it is equally likely that the number of additional novels found is equally
likely to be greater or less than this value?

3. Suppose you were told that the number of additional novels ($n_y$) found in
$y$ is less than $M$; can you now determine a new value such that you feel
it is equally likely that the number of additional novels found is equally
likely to be greater or less than this value?

Answers:

1886 -- currently 305, median is 550, 700 is 75%, 450 is 25% (bassett estimates 10% exclusion)
1891 -- currently 337, 600 is total median, 700 is 75%, 500 is 25% (bassett estimates 10-15% of his novels would be excluded by GRS)
1894 -- currently 385, 900 median, 1100 is 75%, 750 is 25% (bassett estimates 10% exclusion)
